{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.0,
  "eval_steps": 1000,
  "global_step": 16760,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0002983293556085919,
      "grad_norm": 6.0359320640563965,
      "learning_rate": 0.0,
      "loss": 2.2793,
      "step": 1
    },
    {
      "epoch": 0.29832935560859186,
      "grad_norm": 3.87988543510437,
      "learning_rate": 5.960620525059666e-06,
      "loss": 2.265,
      "step": 1000
    },
    {
      "epoch": 0.29832935560859186,
      "eval_cosine_accuracy": 0.9628772735595703,
      "eval_loss": 0.8168219923973083,
      "eval_runtime": 23.679,
      "eval_samples_per_second": 401.579,
      "eval_steps_per_second": 1.605,
      "step": 1000
    },
    {
      "epoch": 0.5966587112171837,
      "grad_norm": 3.6651225090026855,
      "learning_rate": 1.1927207637231503e-05,
      "loss": 1.8687,
      "step": 2000
    },
    {
      "epoch": 0.5966587112171837,
      "eval_cosine_accuracy": 0.9732884764671326,
      "eval_loss": 0.7446231842041016,
      "eval_runtime": 23.5526,
      "eval_samples_per_second": 403.734,
      "eval_steps_per_second": 1.613,
      "step": 2000
    },
    {
      "epoch": 0.8949880668257757,
      "grad_norm": 3.408712863922119,
      "learning_rate": 1.7893794749403345e-05,
      "loss": 1.6653,
      "step": 3000
    },
    {
      "epoch": 0.8949880668257757,
      "eval_cosine_accuracy": 0.9795982837677002,
      "eval_loss": 0.6802141666412354,
      "eval_runtime": 23.5108,
      "eval_samples_per_second": 404.453,
      "eval_steps_per_second": 1.616,
      "step": 3000
    },
    {
      "epoch": 1.1933174224343674,
      "grad_norm": 3.534438133239746,
      "learning_rate": 1.903639618138425e-05,
      "loss": 1.4221,
      "step": 4000
    },
    {
      "epoch": 1.1933174224343674,
      "eval_cosine_accuracy": 0.9819118976593018,
      "eval_loss": 0.6609337329864502,
      "eval_runtime": 23.7912,
      "eval_samples_per_second": 399.686,
      "eval_steps_per_second": 1.597,
      "step": 4000
    },
    {
      "epoch": 1.4916467780429594,
      "grad_norm": 3.8688464164733887,
      "learning_rate": 1.754474940334129e-05,
      "loss": 1.4373,
      "step": 5000
    },
    {
      "epoch": 1.4916467780429594,
      "eval_cosine_accuracy": 0.9834893345832825,
      "eval_loss": 0.633538007736206,
      "eval_runtime": 23.5674,
      "eval_samples_per_second": 403.481,
      "eval_steps_per_second": 1.612,
      "step": 5000
    },
    {
      "epoch": 1.7899761336515514,
      "grad_norm": 3.5698401927948,
      "learning_rate": 1.6053102625298332e-05,
      "loss": 1.3735,
      "step": 6000
    },
    {
      "epoch": 1.7899761336515514,
      "eval_cosine_accuracy": 0.984015166759491,
      "eval_loss": 0.606587827205658,
      "eval_runtime": 24.689,
      "eval_samples_per_second": 385.151,
      "eval_steps_per_second": 1.539,
      "step": 6000
    },
    {
      "epoch": 2.088305489260143,
      "grad_norm": 3.6810312271118164,
      "learning_rate": 1.4562947494033414e-05,
      "loss": 1.2183,
      "step": 7000
    },
    {
      "epoch": 2.088305489260143,
      "eval_cosine_accuracy": 0.9845409393310547,
      "eval_loss": 0.6027300357818604,
      "eval_runtime": 23.7089,
      "eval_samples_per_second": 401.074,
      "eval_steps_per_second": 1.603,
      "step": 7000
    },
    {
      "epoch": 2.386634844868735,
      "grad_norm": 3.416872978210449,
      "learning_rate": 1.3071300715990453e-05,
      "loss": 1.2665,
      "step": 8000
    },
    {
      "epoch": 2.386634844868735,
      "eval_cosine_accuracy": 0.9850667715072632,
      "eval_loss": 0.5950580835342407,
      "eval_runtime": 23.4983,
      "eval_samples_per_second": 404.668,
      "eval_steps_per_second": 1.617,
      "step": 8000
    },
    {
      "epoch": 2.684964200477327,
      "grad_norm": 3.47507905960083,
      "learning_rate": 1.1579653937947496e-05,
      "loss": 1.2465,
      "step": 9000
    },
    {
      "epoch": 2.684964200477327,
      "eval_cosine_accuracy": 0.9855926036834717,
      "eval_loss": 0.5876944065093994,
      "eval_runtime": 23.9275,
      "eval_samples_per_second": 397.409,
      "eval_steps_per_second": 1.588,
      "step": 9000
    },
    {
      "epoch": 2.983293556085919,
      "grad_norm": 19.212100982666016,
      "learning_rate": 1.0089498806682578e-05,
      "loss": 1.1538,
      "step": 10000
    },
    {
      "epoch": 2.983293556085919,
      "eval_cosine_accuracy": 0.985802948474884,
      "eval_loss": 0.6058432459831238,
      "eval_runtime": 23.5551,
      "eval_samples_per_second": 403.691,
      "eval_steps_per_second": 1.613,
      "step": 10000
    },
    {
      "epoch": 3.2816229116945106,
      "grad_norm": 3.6825013160705566,
      "learning_rate": 8.597852028639619e-06,
      "loss": 1.1729,
      "step": 11000
    },
    {
      "epoch": 3.2816229116945106,
      "eval_cosine_accuracy": 0.9856977462768555,
      "eval_loss": 0.5817733407020569,
      "eval_runtime": 25.0755,
      "eval_samples_per_second": 379.215,
      "eval_steps_per_second": 1.515,
      "step": 11000
    },
    {
      "epoch": 3.579952267303103,
      "grad_norm": 3.5933284759521484,
      "learning_rate": 7.106205250596659e-06,
      "loss": 1.177,
      "step": 12000
    },
    {
      "epoch": 3.579952267303103,
      "eval_cosine_accuracy": 0.9863287210464478,
      "eval_loss": 0.5787297487258911,
      "eval_runtime": 23.6358,
      "eval_samples_per_second": 402.314,
      "eval_steps_per_second": 1.608,
      "step": 12000
    },
    {
      "epoch": 3.8782816229116945,
      "grad_norm": 3.369701385498047,
      "learning_rate": 5.614558472553699e-06,
      "loss": 1.1622,
      "step": 13000
    },
    {
      "epoch": 3.8782816229116945,
      "eval_cosine_accuracy": 0.9865390658378601,
      "eval_loss": 0.5771661996841431,
      "eval_runtime": 23.6208,
      "eval_samples_per_second": 402.57,
      "eval_steps_per_second": 1.609,
      "step": 13000
    },
    {
      "epoch": 4.176610978520286,
      "grad_norm": 3.608543872833252,
      "learning_rate": 4.124403341288783e-06,
      "loss": 1.0635,
      "step": 14000
    },
    {
      "epoch": 4.176610978520286,
      "eval_cosine_accuracy": 0.9864339232444763,
      "eval_loss": 0.577176034450531,
      "eval_runtime": 23.9064,
      "eval_samples_per_second": 397.76,
      "eval_steps_per_second": 1.59,
      "step": 14000
    },
    {
      "epoch": 4.4749403341288785,
      "grad_norm": 3.4749655723571777,
      "learning_rate": 2.6327565632458235e-06,
      "loss": 1.1372,
      "step": 15000
    },
    {
      "epoch": 4.4749403341288785,
      "eval_cosine_accuracy": 0.9865390658378601,
      "eval_loss": 0.575279712677002,
      "eval_runtime": 23.9765,
      "eval_samples_per_second": 396.597,
      "eval_steps_per_second": 1.585,
      "step": 15000
    },
    {
      "epoch": 4.77326968973747,
      "grad_norm": 3.465466260910034,
      "learning_rate": 1.1411097852028642e-06,
      "loss": 1.1376,
      "step": 16000
    },
    {
      "epoch": 4.77326968973747,
      "eval_cosine_accuracy": 0.9867494106292725,
      "eval_loss": 0.5734794735908508,
      "eval_runtime": 24.8205,
      "eval_samples_per_second": 383.111,
      "eval_steps_per_second": 1.531,
      "step": 16000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 16760,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}